Bobrocket Posted July 9, 2015 Posted July 9, 2015 Actually it doesn't use the same template for greetings, there is a selection that it randomly chooses from. The screenshot in the post just happens to feature the same template chosen twice in a row. You're right though, it is risky because the standard of the conversations can vary a lot. I had this one recently which was pretty funny, and also fairly human like: Me: Hi X. X: Hello. Me: Hi. X: What's up. Me: Nothing, you? X: Just getting some levels. Me: Getting what? X: Strength. Me: Well in life you must learn to over come trails by believing that you will make it. X: Im too high for that much Me: I don't think s. Then again, you get the occasional awful one like this: X: Norma are u obby maulker? Me: I agree. X: So obby maulers are not 28lv anymore?? Me: I was never Roboton... X: With 60str Me: Your best friend. Your point about using people's full names just gave me an idea. Perhaps it could attempt to split up people's names into sub-names (so "DingleBert3" becomes "dingle" and "bert") and then use those instead. I'll definitely try and implement this An interesting idea definitely! Have you heard of the concept of neural networks? Because what you're suggesting is basically the same principle. Actually implementing this would be seriously hard, partly because you would need to gather a large amount of input data. It wouldn't be hard to gather a large amount of chat messages but gathering logs of people's conversations is basically impossible because the bot would need a near-human level of language understanding to be able to tell when two people are in conversation. If you can't gather conversations then the bot will be able to say runescape-esque things like "buying gf" or "selling lobs 300ea cause i'm an extortionate mf" but won't have any clue how to respond to people in a sensible way. tl;dr Your suggestion is great but would probably take a team of AI professors to implement. You wouldn't necessarily need neural networks, for example: You log persons a and b. This is an extract from their conversation: a: "glastonbury was really insane, i saw so much cool stuff" b: "really? i wish i went!" You can split person a's response like so: {glastonbury} (was|is|has been) (really|very|somewhat) (insane|cool|thrilling), i (saw|viewed) so much (cool|wicked|awesome) (stuff|items) Legend: {} - unnecessary filler words (nouns) (|) - synonyms That way, you can cover sentences like: college has been very cool, i saw so much awesome stuff (I can't think of any more) This isn't really a neural network, because you don't tweak the "nodes" per say. The only learning involved would be some sort of database that lists sentences. Neural networks are a very cool concept indeed, and yeah would need an entire team, but it doesn't need to be like that.
Normangorman Posted July 10, 2015 Author Posted July 10, 2015 You wouldn't necessarily need neural networks, for example: You log persons a and b. This is an extract from their conversation: a: "glastonbury was really insane, i saw so much cool stuff" b: "really? i wish i went!" You can split person a's response like so: {glastonbury} (was|is|has been) (really|very|somewhat) (insane|cool|thrilling), i (saw|viewed) so much (cool|wicked|awesome) (stuff|items) Legend: {} - unnecessary filler words (nouns) (|) - synonyms That way, you can cover sentences like: college has been very cool, i saw so much awesome stuff (I can't think of any more) This isn't really a neural network, because you don't tweak the "nodes" per say. The only learning involved would be some sort of database that lists sentences. Neural networks are a very cool concept indeed, and yeah would need an entire team, but it doesn't need to be like that. Ok, yeah I understand what you mean. But still, there's a difference between saying sentences and having a conversation. The problem remains - how do we detect whether two people are conversing in order to collect data from their conversation to build this database. I guess if there are only 2 people (and you) and they are talking, then they must be talking to each other - but if you think about somewhere like the Grand Exchange, there are hundreds of people all talking at once.
Deceiver Posted July 10, 2015 Posted July 10, 2015 Nice script, but it does get stuck in the North-Eastern corner pen of the cows (ones across the chickens). I just restarted the script but if it happens again, I'll look screenshot it. I think it may have been because the desired skill reached the level I picked.
SocialGFX Posted July 23, 2015 Posted July 23, 2015 Possibly a random, small percentage to misspell words?
sohail Posted August 3, 2015 Posted August 3, 2015 it doesnt work, it just stands there doing nothing 2
datmsbspec Posted September 11, 2015 Posted September 11, 2015 it doesnt work, it just stands there doing nothing this.
randomsht Posted November 29, 2015 Posted November 29, 2015 thank u very much for this, im def gonna use/buy it in the future if needed
